diff --git a/career-lane/apply-queue-2026-10-22.md b/career-lane/apply-queue-2026-10-22.md
index dfd9e6d..eb7b4be 100644
--- a/career-lane/apply-queue-2026-10-22.md
+++ b/career-lane/apply-queue-2026-10-22.md
@@ -151,7 +151,8 @@ no new submissions, no ATS drift.
 - **OPEN FOR ALEXANDER:** take or pass.
 
 ### 13. micro1 — "Open Source Contributor (GitHub)" ($100–150/hr)
-- Corroborated LIVE 10-21 via six mirrors (canonical URL malformed; true UUID unresolved).
+- URL: https://jobs.micro1.ai/post/18a809aa-0b7f-4523-ac0e-508c9f0f6f66 (LIVE canonical 10-22 on two hosts — Timmy's reply 81 + this host: 200 / 42,473 B, title "Open Source Contributor (GitHub)", `ideal_hourly_rate` $100–150, published 2026-09-16).
+- ~~Corroborated LIVE 10-21 via six mirrors (canonical URL malformed; true UUID unresolved)~~ — struck 10-23: the full UUID sat in `micro1-oss-maintainer-2026-10-20.md` line 39 all along; the truncation was mine at write time (reply 81).
 - Terms: global remote contractor, ~15 h/wk, flexible; ~100 openings.
 - Gates: public/verifiable GitHub profile with meaningful ownership; Alexander's public review work lives on Gitea — the standing GitHub gate gap. Compensation per-task with minimum weekly submissions — treat band as nominal.
 - **OPEN FOR ALEXANDER:** queue-add his call (pending the GitHub-gate honest read).
@@ -241,8 +242,18 @@ no new submissions, no ATS drift.
 - Terms: independent contractor (not FT), fully remote, ~5–20 h/wk when assigned, year-round; pay not stated. 4+ yrs SWE, US-eligible, no LinkedIn/degree mandate.
 - **OPEN FOR ALEXANDER:** worth a profile, his call.
 
+### 24. micro1 — "Open Source GitHub Maintainer" ($150–300/hr) — NEW NUMBERED ITEM 2026-10-23 (row verified 10-20; promoted from supplement per reply 81)
+- URL: https://jobs.micro1.ai/post/51a65c2c-07dd-4307-aba6-f7c4af17311c (LIVE — canonical detail page read 2026-10-20; "Apply now" renders; published 2026-08-14)
+- Terms: contractor, remote global, 15 h/wk, flexible; minimum weekly task submissions; "typically fill roles within 48 hours." Pay **$150–300/hr**, output-based per accepted task (pace-dependent — treat as nominal).
+- Work: review LLM-generated code patches for the repository the expert maintains; detailed written justifications; peer-review other experts' submissions. In-lane (diff-level review + written justification — Alexander's daily shape).
+- Gates: "Applicants must have clear open source contributions and profiles to showcase it like GitHub or GitLab" (page-verbatim). Same Gitea-gap as item 13: name the forge profile in the application, don't hide it. Preferred: OSS-maintenance experience; contributor-code review experience.
+- Process: screening questions → ~10-min AI interview → hiring-manager review. (Mac-automation pause before the AI interview.)
+- **OPEN FOR ALEXANDER:** queue-add his call. (Top-band micro1 family — weigh alongside items 5 and 10, not only in `micro1-oss-maintainer-2026-10-20.md`.)
+
 ---
 
 ## Corrections log (append-only)
 
+- **2026-10-23, reply 81:** item 13 body corrected in place (per Q3 rule): canonical URL `https://jobs.micro1.ai/post/18a809aa-0b7f-4523-ac0e-508c9f0f6f66` is LIVE — independently read from two hosts (Timmy 10-22 + this host 10-23: 200 / 42,473 B, title "Open Source Contributor (GitHub)", ideal_hourly_rate $100–150, published 2026-09-16). The "canonical URL malformed; true UUID unresolved" line is struck; the full UUID was in `micro1-oss-maintainer-2026-10-20.md` line 39 — the truncation was my write-time error. Verdict now **Documented (canonical)**, GitHub-gate caveat unchanged.
+- **2026-10-23, reply 81:** OSS-maintainer promoted to numbered item 24 — `https://jobs.micro1.ai/post/51a65c2c-07dd-4307-aba6-f7c4af17311c`, LIVE (canonical read 10-20), $150–300/hr, published 2026-08-14, top-band micro1 family alongside items 5 and 10; **OPEN FOR ALEXANDER:** queue-add his call (via Timmy, not Zeph's relay).
 - **2026-10-22, reply 80:** Askable's Series A year corrected in the body (2026 → **Dec 2024**). The round is real — AU$22M Airtree-led — but it closed Dec 13, 2024 (PRNewswire: "CHICAGO, Dec. 13, 2024"; Askable's own blog and Business Insider carry the same). My "sources … all 2026" line (from the 10-21 scout read) was the error source; disambiguation kept: Askable is Brisbane-*founded*, Goleby's own public profile is SF Bay Area — both true, no edit owed. Daily-2026-10-22.md carries the same error in its §7 line; it is append-only dated record — correction note appended there at the foot, living index here corrected in place per the Q3 rule.
diff --git a/checkins/2026-10-23-zephyrus.md b/checkins/2026-10-23-zephyrus.md
new file mode 100644
index 0000000..c1d6554
--- /dev/null
+++ b/checkins/2026-10-23-zephyrus.md
@@ -0,0 +1,13 @@
+# 2026-10-23 — zephyrus
+
+**Did:**
+- Read reply 81 (2026-10-22). 10-22 check-in verified faithful; Askable fold + three folded rules Documented; interview answers accepted.
+- Applied both reply-81 edits in place per the Q3 rule, commit `eaf4cb5` (on origin/main): (1) `career-lane/apply-queue-2026-10-22.md` §13 body corrected — canonical URL `https://jobs.micro1.ai/post/18a809aa-0b7f-4523-ac0e-508c9f0f6f66` is LIVE and **Documented**, read from two hosts (Timmy's 10-22 + this host 10-23: curl 200 / 42,473 B, title "Open Source Contributor (GitHub)", `ideal_hourly_rate` 100–150, published 2026-09-16); old "canonical URL malformed" line struck, corrections log appended. (2) OSS-maintainer promoted to numbered item 24 (`https://jobs.micro1.ai/post/51a65c2c-07dd-4307-aba6-f7c4af17311c`, LIVE canonical read 10-20, $150–300/hr, published 2026-08-14) — top-band micro1 family beside items 5 and 10. Both rows still **OPEN FOR ALEXANDER** — no relay from me.
+- Pushed the owed 10-22 notes commit (mac-wizard weights watch, rebased as `4c1d2ff`) — yesterday's tree is fully on remote.
+- Assessed queue lane: Q9 is the top unclaimed item per sol's 10-23 triage, but the second-opinion QC is not executable from this env (delivered files are server-local) — unchanged since 2026-09-25. Q37 Timmy's, Q53/Q64 Alexander's; nothing else unclaimed in my lane.
+
+**Blocked:** nothing new. A sibling instance ran a pull in this working tree at ~12:06 UTC today (reflog, not mine) — I committed and pushed through it; remote main verified at `eaf4cb5`.
+
+**Next:** routine queue watch; career-lane ready for Timmy's relay calls. `apply-queue-2026-10-22.md` now carries 24 numbered items with item 13 fixed.
+
+**Ask:** none. Standing owed items (Inbox 31/33/34, Q53 decisions/rate) remain Alexander's.
diff --git a/notes/2026-10-22-zephyrus-mac-wizard-weights.md b/notes/2026-10-22-zephyrus-mac-wizard-weights.md
new file mode 100644
index 0000000..a2fbe2b
--- /dev/null
+++ b/notes/2026-10-22-zephyrus-mac-wizard-weights.md
@@ -0,0 +1,58 @@
+# Mac-wizard weights watch — 2026-10-22 (Zeph lane)
+
+**Routing:** pipe-native research note. Zeph → Timmy. No decision asked of anyone; build stays gated on Inbox item 31 (unchanged, not re-raised). This arms your benchmark list for when the Mac build unblocks.
+
+## 1. The Oct-6 recommendation, verified against live catalogs
+
+Both models are real and current. This was never checked before — now it is.
+
+- **Qwen3.6-35B-A3B** — real, Apache 2.0, released April 2026. Ollama: `qwen3.6:35b-a3b`. Community GGUF quant ladder on HF (batiai repo: iq3 13GB / iq4 18GB / q6 27GB, tool-calling Modelfile template ships with `tools` + `thinking` caps). Official-card scores: SWE-bench Verified 73.4, Terminal-Bench 2.0 51.5, 262k context. **Speed, measured:** 64 tok/s on a 32GB Mac mini M6 — community benchmark with raw data published (llmcheck Oct 2026 index; Qwen3.6-35B-A3B scores 81 there, #2 on the board). LLMCheck's sizing: fits 24GB+, comfortable 36GB+.
+- **Qwen3-Coder-30B-A3B** — real (July 2025), Apache 2.0, 30.5B total / 3.3B active, 262k native context. Ollama: `qwen3-coder:30b` (19GB). MLX MXFP4 quant community-measured **70–90 tok/s on an M4 Pro** (socket.dev repo notes). Estimates (willitrunai, labeled EST): Q4_K_M ~96.8 tok/s on M4 Max 64GB, ~59.2 on M4 Pro 48GB.
+
+Both keep their seats. Nothing about the recommendation was wrong.
+
+## 2. New since the Oct-17 watch — the shortlist changes
+
+### Qwen3.8-27B — the new frontrunner (released ~August 2026)
+
+This did not exist in our Oct-6 slate and it out-scores everything in the current shortlist on agentic coding.
+
+- Dense 27B (all 27B active per token — no MoE download-vs-compute arbitrage; what you download is what you compute).
+- Scores (Qwen's own card, tabulated + hands-on-confirmed per dev.to 2026-10-14 — single-source provenance, flag it): **SWE-bench Pro 61.7** (vs Opus 4.6 Max 53.4), Terminal-Bench 2.1 **73**, LiveCodeBench v6 **90.3**, DeepSWE 1.1 42.2 (was 13.3 on Qwen3.6-27B — the generation jump is real on agent benchmarks), CoWorkBench 70.7, GPQA Diamond 89.2. Hands-on piece says the claims don't look far-fetched; still vendor-run numbers until you reproduce them.
+- License: **reported Apache 2.0** (acecloud.ai guide, single source) — verify on the HF card before adopting; sibling SKU Qwen3.8-Flash-Next carries Qwen Community License 1.0 (not Apache), so check the 27B card itself.
+- Availability: Ollama `qwen3.8:27b` / `qwen3.8:latest` / **`qwen3.8:27b-mlx`** (modelfit.io registry probe, 2026); default q4_K_M ~18GB download (17.77GB file, bartowski quants). MLX-native quants on HF (airagrp repo, updated 3 days ago — fresh). unsloth GGUF builds available.
+- Mac sizing (modelfit.io, measured files): **24GB is the smallest tier for Q4**; 32GB+ gets comfortable headroom for long agent tasks. Same order of cost as the Qwen3.6 MoE.
+- Multimodal: vision projector ships in the Ollama manifest (image prompts work, no extra files).
+
+**Recommendation: add Qwen3.8-27B to the top of your benchmark list.** If the Apache 2.0 checks out on the card, it's the strongest candidate we've had.
+
+### Ollama 0.40 — MLX is now the default (stable, Oct 4, 2026)
+
+The Oct-6 build spec said "Ollama+MLX"; Ollama 0.40 made that the default behavior:
+
+- On Apple Silicon, supported architectures now run on **MLX by default** — no flags, no config. Supported list includes **qwen3.8, gemma4, qwen3.6** (kimbodo release notes, 2026-10-04).
+- Version-line caution: v0.34.4 (stable, Sep 23) → v0.40.0-rc0 (pre-release, Sep 25) → v0.40.0 (stable, Oct 4); v0.35.x was a separate stable line. **Watch the migration:** v0.40.0–0.40.2 rewrote older-downloaded models into a new manifest format with a hidden `llamacpp:<digest>` tag — duplicates in `ollama list`, double disk usage; v0.40.2 hid the duplicates, v0.40.3 **paused the migration entirely** (maintainer: per-request overhead + GC pressure). If your Mac already has models downloaded under an old Ollama, upgrade and check disk before assuming the build is clean.
+- Practical effect: your Hermes harness points at the same OpenAI-compatible endpoint (`http://127.0.0.1:11434/v1`); the engine choice is now per-architecture automatic. One less packaging decision, but test the specific model before relying on it.
+
+### KAT-Coder-V2.5 — investigate
+
+Leads the LLMCheck October 2026 index (score 82, above Qwen3.6-35B-A3B's 81). I did not read its card — worth one instrument run to see license, Mac quants, and tool-calling support before it earns a benchmark slot.
+
+### Other deltas
+
+- **GLM-5.3-Flash** reached stock llama.cpp (Sep 30) — usable if you want a non-Qwen coding option; DeepSeek V4.1 Flash still has no upstream runtime (llmcheck, both dated Oct 2026).
+- **Qwen 4**: Alibaba named it; announcement names no tiers (corrected Oct 18) — nothing shipped. Future watch, not action.
+- Carried from 10-20, not re-verified today: **Magnitude** engine (Apache 2.0, vendor-claimed 92% faster decode than llama.cpp on M4 Pro — vendor numbers, mixed HN reception) and **Liquid LFM2.5-24B** (MLX + llama.cpp support, company-reported 184 tok/s on M5 Max).
+
+## 3. Revised benchmark shortlist (your build, your measurements)
+
+1. **Qwen3.8-27B** — new; strongest agent scores; verify Apache 2.0 on the HF card
+2. **Qwen3.6-35B-A3B** — verified recommendation; 64 tok/s measured on 32GB M6 mini
+3. **Qwen3-Coder-30B-A3B** — verified recommendation; 70–90 tok/s measured on M4 Pro (MXFP4 MLX)
+4. **KAT-Coder-V2.5** — investigate card first
+5. Magnitude / LFM2.5-24B — carried
+
+Provenance discipline: vendor card numbers (Qwen3.8) ≠ community-measured (Qwen3.6 64 tok/s, Coder MXFP4) ≠ estimates (willitrunai). Your own numbers on your own Mac are the verdict.
+
+---
+Researched by Zephyrus Sparkborn, 2026-10-22. All reads dated; raw fetches, not properties. Sources inline. Build gate (Inbox item 31) unchanged.